Stock Ryzen 5 1600x cannot handle realbench?

dweenjatheninja

Prominent
Jul 28, 2017
32
0
530
Hey all,

I've recently just built my very first PC and after reinstalling windows because I didn't make my drive a partition when installing windows, I thought I would pay stability a little attention. I got the 1600x over the 1600 because I thought that meant I would be able to use it at 3.6GHz out of the box just fine. Which I do, but at the cost of VCore going up to 1.55, so not that great. To my disappointment, I had to find a way to decrease Voltage so I played around in the BIOS, where i realized my RAM was at only 2133MHz as opposed to the native 2666MHz. Changing this to DOCP 2666MHz profile with a matching timing automatically set, it fails to POST. But cool cool, skip this and learn how to offset VCore(kinda proud of myself btw haha, idiot taking baby steps at learning to #pcmasterrace).

Offset by 0.15V with regular LLC and system crashes. So go back to BIOS and change offset to -0.1V and set LLC to Medium and all is good in the hood. Try some benchmarks when paired with my 980TI, get impressed and try a stress test. Turn on realbench and it wont even run for more than 10 seconds before declaring "...instability (blender)".

So after all this, apart from worrying about the health of my PC(cause after the crash my SSD was not detected in boot), i decided to revert BIOS settings to default and try the stress test again. The stress test still faces the same issues trying to run at 8GB RAM capacity(am using 2 x 8GB kit). What am I doing wrong? cause i was so happy when the system was running basic surfing stuff at the voltage offset.

Does passing stresstests matter? ): I haven't got a monitor yet so the past few days have been trying to tweak my PC to perfection before I dive in to hardware heaven. Really guys, any help would be great now cause sorry to say, but at this point I really know nothing.
 
Ok, I will need some specs.

What ram did you buy? CPU cooler?
You doing stock 1600X or trying to overclock? I was confused about your 1.55 vCore setting if you are not overclocking.

Stability is important especially when you are just running basic programs and getting it to crash.
 


Heya! thanks for getting back to me so fast! The 1.55V is part if the PB/XFR I guess. My RAM is Vengeance LED 2666MHz 2x8GB, The 2x16GB is listed in the QVL so I stupidly assumed that the 2x8GB would work fine too. MOBO is Strix B350-F by the way. Cooler is the H7 Quad Lumi, temps seem fine(i think), at ~45-52 at idle. Haven't got to successfully stress test or game so don't know at load. :/

As for the voltage, its not constant at 1.55V but it spikes up to there. On idle it stays at around 0.83 but will spike up randomly, legit all the way up to 1.55v. Got it on both Ryzen Master and CPU-Z. Haven't really gamed to test anything at the moment. But the Cinebench results seem pretty good when I was offsetting by -0.1V. By the way when I offset, I still left PB/ZFR on.

I guess the major concern is that when I tried the realbench at default BIOS setting it still presented me with the same instability message. Really need help man! ):
 
OK.

I think with the quad lumi your idle temperatures should be way lower than that. I would say 35-40C. I would clean and reapply paste and remount the cooler.

Did you flash to the newest BIOS for your mobo? Ryzen is VERY picky about ram and BIOS updates are very important to get the XMP (DOCP) profiles working.

1.55V is VERY high for a 1600X at full load at the auto voltage setting. I'm surprised ASUS would allow that. I would flash the BIOS if you didn't.

I think your crashing issues is with cooling mostly. Install HWinfo and see what the package temperature is at full load and let me know.

------------
I just built an overclocked 1600X system so I went through all of this. Not the same parts though.
 


I had HWInfo64 running before I reformatted my drives, and found that the temperatures have a 20C offset and noted that Ryzen Master actually reports the actual temperature before offset so my reported temperatures have nothing to do with that. :/

As for the BIOS, it is up to date for me at version 0806, tried to change to DOCP profile after update but crashed like before, 2-3 POST attempts and then booting into safe BIOS prompting me to revert to default RAM settings. Is it cause the 1.2V dram voltage is too low?

Guessing the temperature is mostly due to the high voltage but as I've mentioned, trying to offset the voltage by -0.1 with LLC at Medium(no system crashing(here I would be seeing idle vcore at 0.71 but peaks at ~1.45, might step up LLC to High instead for a more idle Vcore)) still fails realbench after 10 seconds.

By the way, any thoughts on a -1.25~1.5V offset with Extreme LLC and PB/XFR enabled? Or is that a no-no?

Or is this the silicon lottery everybody is talking about? I already sent a refund request to the guy who sold me the CPU cause the 212 LED mounting mechanism on AM4 is a bit meh lol. So trying to avoid troubling him even further.
 
So can you do one thing at a time? Just let your memory stay at the default, put everything on auto. Your system should be stable at this point. If your idle temperatures at this point is 45-50C that is too high. If you system is not stable doing stability test with all default settings then something is wrong with your system but you don't believe me that the temperature is too high at full load so :/.

Silicon lottery is for high overclocking people. Your system should function at stock so it's not silicon lottery.
 


I do believe you. ): sorry i was just trying to avoid reapplying my stock thermal paste cause I dont have anything to wipe it down with. :/ but okay I will give that a shot
 


You will need 99% isopropyl alcohol and some paper towels.
 


Make sure your bios is updated. The latest bios versions have the updated AGESA code, that allows for better memory overclocking. The earlier versions of the bios were really poor for memory.

The early versions of the bios were also known to set voltages way too high when left on "auto".

 
Hello there, thanks for trying to help too, just to update you guys, last night since i didn't have any isopropyl i couldn't clean the heatsink so i've left it as such. BIOS is in fact up to date but memory overclocking still not proving fruitful at all. What has changed is that I've set the following:
VCore Offset -1.1125 LLC High
SOC Voltage +1.1125 LLC Low(or medium, i'm at work i cant remember or check)

I ran Prime95 for about 8 and a half hours, results were 1.22V to maintain 3.7GHz at average temp( for the first hour while i was awake to monitor) was ~62C +- 2C.

New idle temperature ranges from 33-40C, idle voltage or at low load still spikes from 0.7V to 1.43V as per XFR i think cause i didn't set a fixed multiplier. Should I be happy with these results? Should i offset the VCore a bit more and retest or are these results considered okay?

 


8 1/2 hours on P95 is more than enough to consider it stable. As long as your temps stay at a manageable level, no reason to not push for 3.9. Seems like you're getting a handle on your bios.....so, you're already half way there.
 


Thanks for the help though, haven't pushed the clock speed cause i just let the CPU go on turbo on its own accord, but yeah, it does make the voltage spike from 0.7V @2.2GHz up to 1.43V @4GHz on turbo, and this will happen very quickly and last for a second maybe? Is this bad for a CPU is what I am wondering, if so then i'll just fix the multiplier and properly learn how to OC my cpu haha.

I think its cause i think of the CPU as the engine of the car, in which case, it wouldn't be good if i keep the engine idling, then suddenly switch gear and step on the pedal, again and again and again lol. But it is the stock setting, so its not like they CPU was built to destroy itself is it?

Really hoping i can get a handle on this, and that every other pc noobie using the same proc/mobo can put their worries to rest too
 
GOod lord...

Reset the CMOS, enter BIOS, reset to defaults/auto/SPD, etc..

THe jumping to 1.5V on Vcore is downright scary if that happened....

Voltage will absolutely normally spike nearly instantaneously up and down from .7V to 1.35V or so to accommodate higher clockspeeds..

 


1.43v intermittently isn't going to really hurt anything, but.....It really shouldn't need that much voltage to run the stock clock settings.

What motherboard do you have?
 


Hey there! I tried to jump the CMOS, no efefct to my vcore problem. But I can honestly say I am quite comfortable with the settings at the moment. I stepped the negative offset down further to -0.125. For the most part I am seeing idle temps of 32-40 now @ 0.68V and 2.2GHz. When I ran P95 for an hour today, I've got 3.7MHz on all cores @1.198V and 61+-1C. My motherboard is the strix b350f by the way. Did not know there would be so much troubleshooting and tweaking with PC building to be honest, but its cool; the more problems I'm running into the more I'm learning about this. Would probably have never opened my BIOS except to change boot if it wasn't for this whole experience.